Benefits of polyspecific associations for the GoeldiТs monkey (Callimico goeldii).код для вставкиСкачать
American Journal of Primatology 54:143–158 (2001) Benefits of Polyspecific Associations for the Goeldi’s Monkey (Callimico goeldii) LEILA M. PORTER* Interdepartmental Doctoral Program in Anthropological Sciences, State University of New York at Stony Brook, Stony Brook, New York Polyspecific associations are an important component of Callimico goeldii behavior and ecology. On average, Callimico goeldii was found in proximity to or in vocal contact with Saguinus troops (S. fuscicollis and S. labiatus) during 53% of all time intervals sampled. Polyspecific associations varied considerably between seasons, however, with association rates peaking during the wet-season month of February (89%) and declining in the dry season, with the lowest rate (13%) in July. The primary benefits of associations appear to be an increased use of the lower and middle canopy, and an increase in feeding behaviors during the wet season. Thus, Callimico goeldii appear to benefit most from associations during the wet season when fruits are its principal food source. Fruits are eaten more in the forest canopy than in the understory; thus, an increase in height use likely permits an increase in the fruit resources on which Callimico goeldii can forage and feed. In addition, Saguinus groups, with their smaller home ranges, are likely to be more knowledgeable than Callimico goeldii about the location and abundance of ripe fruits in their home ranges. Thus, Callimico goeldii may parasitize Saguinus for their fruit knowledge by following them through their ranges. In the dry season, limited dietary overlap between Callimico goeldii and Saguinus groups is likely to make associations less beneficial for Callimico goeldii as they adopt different foraging and ranging strategies. Am. J. Primatol. 54:143–158, 2001. © 2001 Wiley-Liss, Inc. Key words: Callimico goeldii; Saguinus fuscicollis; Saguinus labiatus; mixed species; competition; predation; foraging; home range INTRODUCTION Among the unusual characteristics of the neotropical subfamily Callitirichinae, the marmosets and tamarins, is their tendency to form stable, long-lasting poly- Contract grant sponsor: Fulbright Scholarship; Contract grant sponsor: National Science Foundation; Contract grant number: 9815171; Contract grant sponsor: Chicago Zoological Society; Contract grant sponsor: LSB Leakey Foundation; Contract grant sponsor: Douroucouli Foundation; Contract grant sponsor: Primate Conservation, Inc.; Contract grant sponsor: Margot Marsh Biodiversity Foundation. L.M. Porter’s present address is Department of Conservation Biology, Chicago Zoological Society, Brookfield, IL 60513. *Correspondence to: Leila Porter, Correo Central, Cobija, Departamento de Pando, Bolivia. E-mail: email@example.com Received 1 March 2000; revision accepted 14 March 2001 © 2001 Wiley-Liss, Inc. 144 / Porter specific associations [Yoneda, 1981; Pook & Pook, 1982; Terborgh, 1983; Yoneda, 1984; Garber, 1988; Buchanan-Smith, 1990; Heymann, 1990; Norconk, 1990; Peres, 1992a–c; Lopes & Ferrari, 1994; Buchanan-Smith, 1999]. The first studies of wild Goeldi’s monkeys (Callimico goeldii) showed that they formed polyspecific associations with two tamarin species, the saddle-backed tamarin (Saguinus fuscicollis) and the red-bellied tamarin (Saguinus labiatus) [Pook & Pook, 1982; BuchananSmith, 1991; Christen & Geissmann, 1994]. The frequency and longevity of these associations, however, remained uncertain due to the short duration of these investigations and lack of fully habituated animals. Polyspecific associations are not likely to occur between species with very disparate behavior. In order to travel and forage together, associated species must have some degree of ecological similarity which allows coordination of group activities [Terborgh, 1983]. However, with increasing ecological similarity comes a potential increase in feeding competition. For some species, increasing group size has been shown to increase feeding competition between individuals within a single-species group [van Schaik & van Noordwijk, 1986; Janson, 1988; Janson & Goldsmith, 1995]. Similarly, polyspecific associations increase overall group size and may increase feeding competition among individuals of a polyspecific group if dietary overlap is high for spatially restricted resources and/or where knowledge of resource location is critical to feeding success [Podolosky, 1990; Terborgh, 1990]. Thus, polyspecific associations are likely to represent a compromise between competition and compatibility, but the benefits should outweigh any potential costs incurred through increased feeding competition. Polyspecific associations are found in a variety of other animals, including ungulates [Fitzgibbon, 1990], birds [Munn & Terborgh, 1979; Popp, 1988], fish [Landeau & Terborgh, 1986], and spiders [Hodge & Uets, 1996]. No single factor can universally explain why these associations occur, as differences between species living in closed and open habitats, the types of predators present, and resource distribution all appear to influence association patterns and their formation [Cords, 1990a; Fitzgibbon, 1990; Terborgh, 1990; Chapman & Chapman, 1996]. This work examines four hypotheses to explain the occurrence of polyspecific associations among Callimico goeldii and sympatric species of Saguinus. First, associations may represent chance encounters between different groups of species sharing a common home range [Waser, 1982, 1984; Whitesides, 1989], and/or common food resources. Second, polyspecific associations may improve predator avoidance. Predation is considered to be an important selective force on the small-bodied callitrichines, which are thought to be prey to a variety of mammalian, avian, and reptilian predators [Peres, 1993]. Although published accounts of predation on callitrichines are scarce, there is increasing evidence that primates are an important part of the diet of terrestrial and aerial predators [Boinski, 1987; Isbell, 1990; Wright, 1998]. Some studies of African primates support the antipredation hypothesis, showing that polyspecific associations increase in frequency when predation risks are highest [Oates & Whitesides, 1990; Holenweg et al., 1995; Nöe & Bshary, 1997]. Increasing group size through polyspecific associations can result in several potential antipredation benefits. Large polyspecific troops may reduce the probability that any one individual in the troop will be preyed upon, thereby reducing an individual’s risk of predation [Roberts, 1996]. Predator avoidance may further improve if associated species are active at different heights in the forest, leading to a greater spread of vigilant individuals in an area than are present in a monospecific group. With greater vigilance spread the troop can more rapidly Callimico goeldii Polyspecific Associations / 145 detect predators [Pook & Pook, 1982; Peres, 1989; Peres, 1992b; Caine, 1993; Peres, 1993; Hardie & Buchanan-Smith, 1997]. Callimico goeldii, Saguinus fuscicollis, and Saguinus labiatus travel at different heights in the forest [Porter, 2000], making it possible for improved detection of predators during associations through greater vigilance spread. Thus, if polyspecific associations improve predator avoidance, Callimico goeldii is predicted to have increased survival rates while associated with Saguinus troops. This, however, can only be measured with longterm data on group demographics and predation rates. It is possible, however, to assess whether associations improve predator avoidance through other measures. The group size effect [Pulliam, 1973] predicts that individuals in large groups can reduce vigilance behaviors because the group as a whole can maintain a high vigilance rate. Callitrichines use scanning to remain vigilant to their predators [Caine, 1984; Ferrari & Lopes Ferrari, 1990; Koenig, 1998], and scanning comprises a large part of the daily activities of callitrichine species [Ferrari & Lopes Ferrari, 1990; Savage et al., 1996; Koenig, 1998]. Thus, a reduction in scanning can lead to an increase in time spent foraging and feeding if these activities are time limited. The group size effect has been supported by studies of polyspecific association among some species [Cords, 1990b; Hardie & Buchanan-Smith, 1997] but not for others [Treves, 1998; Garber & Bicca-Marques, in preparation]. Therefore, if Callimico goeldii benefits from the group size effect it is predicted that it will decrease scanning and increase feeding during associations with Saguinus troops. Third, it has been suggested that some species exploit the resource knowledge of their associates, with one species using another species to lead it to food resources. Saimiri sciureus follow Cebus apella groups around their territories, parasitizing the Cebus groups by following them to fruit resources [Terborgh, 1983; Podolosky, 1990]. Cords [1990b] also found that Cercopithecus ascanius uses Cercopithecus mitus as a guide to food resources in areas where they have comparable diets. If Callimico goeldii is shown to follow at least one Saguinus species, and if Callimico goeldii gain from Saguinus resource knowledge, it is predicted that Callimico goeldii will consume foods more frequently when associated than when alone. Fourth, improved insect foraging may result from the flushing of insects from the middle and upper canopies by one species to an associated species traveling below [Munn & Terborgh, 1979; Terborgh, 1990]. This has been shown to be one advantage that Saguinus fuscicollis gains from associations with S. mystax [Peres, 1992b]. If Callimico goeldii benefits from insect flushing from Saguinus, it is predicted that Callimico goeldii will forage for insects at lower levels than at least one Saguinus species, and have increased insect feeding rates while associated with Saguinus troops than while it is alone. METHODS Study Site Research was conducted in northern Bolivia, in the Department of the Pando, at a field camp at San Sebastian (11° 24′ S, 69° 06′ W, ca. 280 m elevation), 3 km north of the Rio Tahuamanu and 42 km east of the border of Peru [for a detailed description see Porter, 2000; in press]. There are 10 species of primates in the study area, and many species of large predators, including jaguars (Panthera onca), ocelots (Felis concolor), margays (Felis pardalis), tayras (Eira barbara), harpy eagles (Harpia harpjya), and tree boas (Corallus hortulanus). 146 / Porter Study Groups Following a 7-mo habituation period, observations of one group of Callimico goeldii were taken each month from April 1998–March 1999. Group scans on the Callimico goeldii were taken from 7–14 days each month for a total of 957 hr of observations. Due to the large home range size of the Callimico goeldii group, it was observed within the home ranges of seven mixed species (Saguinus fuscicollis and Saguinus labiatus) groups (referred to as groups I–VII). Two of these groups were observed for this study. Observations of Saguinus fuscicollis in group I were taken from April 1998–September 1998 (250 observation hr), and in group II from October 1998–March 1999 (188 observation hr). Observations of Saguinus labiatus in group I were taken during June 1998–September 1998 (123 observation hr) and from group II during October 1998–March 1999 (201 observation hr). Behavioral Methods Instantaneous group scan samples (10 sec in duration) were collected every 5 min [Martin & Bateson, 1993] from the time the groups left their sleep sites in the morning until they retired to a sleep site at night. Animals located only after they moved were not recorded in the scan to avoid biasing the sample towards more active behaviors. Infants were recorded as a separate individual only when moving independently of an adult. For each scan sample the following data were recorded for each individual observed in the group: its general and specific behavior, and the height class and habitat type it currently occupied. Behaviors recorded were foraging (including manual and visual search for food objects), feeding (eating or manipulating a food object), traveling (including quadrupedal movement, leaping, and vertical ascent and descent), resting (including scanning, grooming, and sitting) and other (including play and aggression). Habitat types discussed in this work include: A, primary forest with open understory (canopy > 15 m, visibility at eye level > 20 m); C, primary forest with dense understory (canopy > 15 m, visibility at eye level < 20 m); D, bamboo dominant plant species (canopy < 15 m, visibility at eye level < 20 m); and F, stream’s-edge habitat (canopy > 15 m, visibility < 20 m, dominant plants ferns and palms). Height classes were defined as follows: 0 = 0 m; 1 = 0–5 m; 2 = 5–10 m; 3 = 10–15 m; 4 = 15–25 m; and 5 = >25 m. The group’s location (if known) in the trail system was also indicated during each scan. Interspecific associations, including both physical proximity and vocal contact, were noted using presence–absence sampling between each 5-min scan. Two species were considered to be in physical association if at least one individual of another species was within 15 m of the study group under observation. Although other studies have used greater distance criteria (e.g., 20 m [Cords, 1990b], 25 m [Nöe & Bshary, 1997], and 50 m [Buchanan-Smith, 1990; Chapman & Chapman, 1996; Wachter et al., 1997]), 15 m was chosen for this study as Callimico goeldii was found below 5 m during 80% of observations and in dense understory forest with visibility less than 20 m during 76% of all observations [Porter, 2000]. Thus, a distance criterion of 15 m ensured that association data were recorded accurately, although more strictly than previous studies. Two species were considered to be in vocal contact with one another if contact calls, feeding calls, or alarm calls of another species were audible by the observer following the focal group. The use of presence–absence sampling allowed for the recording of vocal associations, behaviors that are rare and short in duration but Callimico goeldii Polyspecific Associations / 147 very important for group movement and cohesion [Pook & Pook, 1982]. The use of 1:0 sampling should not overestimate association duration [Martin & Bateson, 1993] as the typical association was generally long-lasting and uninterrupted for much longer periods of time than the sample interval used (5 min). Polyspecific group leadership was recorded on 14 days. The leader was defined as the species that physically (for travel or feeding) or temporally (for longcall communication) began an activity that one or more associated species followed. The initiator of an association was defined as the species that moved towards another group, with the result that they traveled to within 15 m of one another. The species initiating the change was noted every time there was a change of activity. Travel was noted as having begun if one or more individuals of a species left its resting, feeding, or foraging site and was then followed by the rest of the group. No leader was noted if there was no species clearly leading the group, or if an activity change was ambiguous. Renkonen’s Percentage Similarity Index was used to calculate diet overlap between species. An index value of 1 indicates complete dietary overlap, and a value of 0 indicates no dietary overlap [for complete description see Porter, in press; Wolda, 1981]. For dietary overlap calculations, fruits, nectar, and exudates were categorized by species, whereas arthropods and fungus were categorized as food types. Group scans were conducted rather than focal animal scans, as individuals of the Callimico goeldii study group were not trapped or marked and therefore were not easily identifiable. Due to the difficulty locating all individuals of a group, the behaviors of each individual recorded (independent records) during one scan were converted to a percentage of the total activity records for that sample. Weighting scans may incur some error as the study groups were not observed completely during each scan and did not have similar proportions of all age classes each month (Table 1). However, as callitrichine groups have been shown to travel and feed in cohesive units [Garber, 1993; Garber, 2000; Peres, 2000], the behavior of one individual closely approximates the behavior of other group members during a scan sample. Thus, for this study, each scan had equal weight in the overall analyses [as in Milton, 1980], and these weighted rates rather than absolute rates were used throughout analyses. TABLE I. Group Composition and Monthly Observation Rates Month April May June July August September October November December January February March Group composition Number of individuals observed/scan 3f, 3m, 1i 3f, 3m, 1i 3f, 3m, 1i 3f, 3m, 1i 2f, 2m, 1j, 1i 2f, 2m, 1j, 1i 2f, 2m, 1j, 1i 2f, 2m, 1j, 1i 2f, 2m, 1j, 1i 1f, 2m 1f, 1m 1f, 1m, 1i 2.72 2.73 2.53 3.14 2.96 3.54 3.67 4.5 4.6 2.6 1.77 1.95 a, adult (unknown sex); f, adult female; m, adult male; i, infant (age 0–6 months); j, juvenile (age 6–12 months). 148 / Porter Analyses Location data recorded each day allowed for an assessment of the area of the home range Callimico goeldii used for that day. The home range was divided into eight areas corresponding to the core areas of the sympatric Saguinus fuscicollis and S. labiatus groups within the Callimico goeldii range. This allowed an assessment of the time Callimico goeldii spent within the range of each Saguinus group and the time it spent in association with each group. To determine whether Callimico goeldii associated with Saguinus species more than expected by chance, I performed the following analysis. I calculated the percentage of observational scans Callimico goeldii occupied 19 (1 ha) plots of forest within Saguinus fuscicollis group I’s home range, and the percentage of observations that S. fuscicollis group I used these same 19 plots. Using these plot-use data, I then calculated the probability the two species would occupy the same plot at the same time (Callimico goeldii % plot use × Saguinus fuscicollis % plot use). This represents the expected percentage of time the species would be together if their proximity were merely chance encounters in a commonly used area. The expected rate was then compared to the actual association rate (based on the much more conservative criteria of 15-m proximity) using a pairwise t-test (SPSS© 9.0 for Windows) to assess whether associations were due to chance alone. All percentage data (behavior, diet, and habitat use) were arcsine transformed before analyses. Using the entire data set, monthly differences in association rates were compared using a standard factorial ANOVA with unbalanced sample sizes (SPSS© 9.0 for Windows). A model 1 step-wise multiple regression (SPSS© 9.0 for Windows) was then used to test which of the following variables best predicted monthly changes in association frequency: group size, dietary overlap, and frequency of feeding, insectivory, frugivory, and mycophagy. For comparisons of Callimico goeldii behavior in and out of association, subsamples of the data were selected from four 1-hr periods (between 8:00–9:00, 10:00–11:00, 12:00–13:00, and 14:00–15:00) across the entire data set to control for possible temporal variation in behaviors, and hourly averages of behaviors were used for analyses. Hours during which Callimico goeldii was alone (100% of all 5-min samples for that hour show no polyspecific associations) were compared to all hours that it was associated (groups were associated with one or more species during at least 10 out of 12 5-min intervals for that hour). In all associated samples a 5- or 10-min lapse in association was immediately followed by at least 1 hour during which the groups were continuously associated; thus these brief lapses were not indicative of a splitting of the polyspecific troop. Interactions between association status (alone or associated) with season (dry season May–October, or wet season November–April) and time of sample were examined for all analyses. Unless otherwise stated, interactions among these fixed factors (time of day, season, or association status) were not statistically significant. To test for differences in height use during sample periods, the KomolgorovSmirnov test [Sokal & Rohlf, 1995] was used to compare the average height used during hours associated vs. during hours alone. The median of each height class was used for calculations (e.g., height class 1 = 2.5 m). RESULTS Null Hypothesis Callimico goeldii was found within Saguinus group I’s home range during 20% of all observations. From these observations, grid location data was avail- Callimico goeldii Polyspecific Associations / 149 able for 47% of scans (n = 1,224). Location data for Saguinus fuscicollis group I were available for 30% of scans (n = 1,145). These location data were used as an estimate of the total frequency that plots were used. Using these frequencies, I calculated the expected percentage of time the two species would occupy any given plot at the same time if associations were due to chance (Callimico goeldii % plot use × Saguinus fuscicollis % plot use) (Table 2). The observed rate was significantly different than the expected rate (df = 18, t = 5.56, P < 0.001); thus, associations were not due to chance alone. Frequency of Associations With Saguinus Association data were examined for all Callimico goeldii observations taken throughout the year. Callimico goeldii was found within 15 m of both Saguinus species at the same time during 22% of observations. Callimico goeldii was found within 15 m of only Saguinus fuscicollis during 22% of observations, and within 15 m of only S. labiatus during 2% of observations. Callimico goeldii was in vocal association with a Saguinus group during 7% of intervals. Callimico goeldii was alone (not within 15 m of, or in vocal contact with either Saguinus species) 47% of the time. In comparison, data from group II shows that Saguinus fuscicollis was in association with S. labiatus during 72% of all records (67% within 15 m and 5% in vocal contact) and with Callimico goeldii during 21% of all records (20% within 15 m and 1% in vocal contact). Frequency of Associations With Different Saguinus Groups Callimico goeldii, unlike Saguinus, do not maintain association with only one group. The home range of the Callimico goeldii study group, with an area of TABLE II. The Observed and Expected Frequency of Association Between Callimico goeldii and Saguinus fuscicollis Within Test Plots Plot Number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Observed Percent 0.13 0.66 0.6 0.1 0.8 0.1 0.6 0.4 0.7 0.21 1 1 0.25 0.03 1 0.47 0.25 0.03 1.4 Expected Percent 0.02 0.13 0.05 0.01 0.14 0.01 0.02 0.02 0.04 0.01 0.09 0.05 0.02 0.02 0.07 0.01 0.00 0.00 0.09 150 / Porter 150 ha, included the entire home ranges of six Saguinus groups, and parts of two other groups (Fig. 1). Callimico goeldii spent the most amount of time within the home range of Saguinus group II, smaller amounts of time with troops I, III, IV, V, and VI, and only brief periods with troops VII and VIII. The percentage of time that Callimico goeldii was associated with Saguinus groups in these areas are indicated in Fig. 1. Associations with a particular Saguinus group (I–VIII) were maintained for as little as half an hour, and as long as several days. Patterns of Group Transfers Three patterns of association transfers were recorded during 40 sequences when Callimico goeldii moved from one Saguinus territory into another. First, Callimico goeldii switched directly from one Saguinus troop to another at the Saguinus troops’ territory boundary (n = 11). Second, Callimico goeldii simply abandoned their Saguinus associates when it was near a Saguinus territory boundary. The Callimico goeldii group would then travel by itself in the new territory, giving long calls intermittently until it encountered a new Saguinus troop on the same day (n = 11) or after an entire day or more alone (n = 7). Third, Callimico goeldii would also leave one Saguinus group at their sleep site in the late afternoon, but then would continue to travel so that in the morning Callimico goeldii began the day with a different Saguinus troop (n = 11). Role of Species Within the Polyspecific Group Callimico goeldii generally initiated associations (66% of records), as it was usually the first species to start contact calls to locate Saguinus groups. All three Fig. 1. The home range of one group of Callimico goeldii (dashed line) in relation to the home ranges of eight Saguinus groups (S. fuscicollis and S. labiatus). The total proportion of time Callimico goeldii spent in each range (TT) and the proportion of that time spent in association with each Saguinus group (TA) is listed below each group number. For simplicity only the core areas of a Saguinus troops—not overlapping boundary areas—are shown. Callimico goeldii Polyspecific Associations / 151 species responded to each other’s contact calls, and this often led the species to form polyspecific groups immediately after they descended from their sleep trees, and allowed coordination of group activities throughout the day. Once the polyspecific group had formed, Saguinus labiatus almost always led group travel and feeding at common food resources, and it was generally the species that initiated resting bouts (Fig. 2). Neither Callimico goeldii nor Saguinus fuscicollis were observed directing any aggression (charging or chasing) at S. labiatus. Callimico goeldii and Saguinus fuscicollis engaged in aggressive interactions on 39% of observation days, but as these episodes were brief; they formed less than 1% of observations during their association. Thus, feeding competition by aggression appeared to be minimal among species. Effects of Polyspecific Association on Callimico goeldii Behavior Habitat use by Callimico goeldii varied with association status and by season. The mean frequency with which stream edge habitats were occupied by Callimico goeldii was higher while alone than while associated (F[1, 180] = 10.64, P < 0.001) and more in the dry season than the wet season (F[1, 180] = 3.92, P < 0.05) [dry season (alone 12%, associated 3%), wet season (alone 5%, associated 0%) (Fig. 3a)]). In contrast, use of primary forest with open understory was highest during the wet season during associations (wet season [associated 8%, alone 0%], dry season [associated 2%, alone 0%], F[3, 178] = 5.58, P < 0.05 (Fig. 3b)). The mean frequency with which Callimico goeldii used bamboo forest (habitat D) was higher when alone (16%) than when in association with Saguinus troops (3%) regardless of season (F[1, 180] = 15.10, P < 0.001). In contrast, the mean frequency with which Callimico goeldii used habitat C (primary forest with dense understory) was higher while Callimico goeldii was associated (88%) than when it was alone (67%) regardless of season (F[1, 180] = 18.68, P < 0.001). Fig. 2. Species initiating mixed group activities for different activities. Cg = Callimico goeldii, Sl = Saguinus labiatus, and Sf = Saguinus fuscicollis. 152 / Porter Fig. 3. Frequency with which Callimico goeldii behavior changed according to association status and season: (a) streamedge habitat use, (b) primary forest with open understory use, and (c) feeding. Error bars show the SEM ± 1.0 SD. Height use varied with association status in habitat C. By using data only from habitat C, I controlled for potential differences in the height at which an animal travels in habitats with different vegetation profiles. Callimico goeldii was found at higher heights (average 3.96 m) while associated than while alone (average 3.14 m) within habitat C (Z = 1.43, P < 0.05). I also examined scanning within habitat C. If data from all height classes are combined, scanning is significantly less when Callimico goeldii is part of a polyspecific group (mean 53%) than when it is alone (mean 62%) (F[1, 98] = 10.77, P < 0.001). If, however, data from only the same height class is used (height class 1: 0–5 m) to control for potential differences in scanning rates in different height classes, there is no significant difference in scanning, although on average Callimico goeldii scanned less during associations (57%) than when alone (65%) (F[1, 29] = 2.26, P < 0.14). Thus, Callimico goeldii remains highly vigilant through scanning within height class 1 regardless of association status. Although not significant, there was a trend toward increased frequency of travel by Callimico goeldii when it was in association (14% of observations) vs. when it was alone (11% of observations) (F[1, 98] = 3.46, P < 0.07). However, as distances traveled were not taken, it is not certain if increased travel frequency corresponds with increased day ranges. Increased travel may be indicative of increased feeding competition within polyspecific troops. Therefore, the effects of associations on feeding behaviors were examined. Feeding frequencies were compared between when Callimico goeldii was alone vs. when it was associated using samples collected only within habitat C. In this way the analysis controlled for potential differences in food availability between habitats. Feeding rates were significantly higher during the wet season while Callimico goeldii was associated (13%) than when it was alone (2%), whereas feeding rates varied little in the dry season between times when it was associ- Callimico goeldii Polyspecific Associations / 153 ated (7%) and when it was alone (8%) (F[3, 96] = 13.31, association × season P < 0.001 (Fig. 3c)). I examined the types of food eaten during samples in habitat C (in which feeding occurred) to determine if associations affected the frequency with which different types of food were consumed. First, I examined whether arthropod foraging was affected by flushing of arthropods by one species to another. Using the height at which arthropods were eaten as an approximation of the height at which they were captured, I found that Saguinus labiatus captured insects at higher heights than the other two species (mean insect eating heights: Callimico goeldii 2.94 m (92% of insect feeding records were below 5m); Saguinus fuscicollis 3.88 m; S. labiatus 11.78 m). Therefore, as Saguinus labiatus is the species that potentially flushes insects to Callimico goeldii, I compared arthropod feeding only during hours when Callimico goeldii was in association with both Saguinus labiatus and Saguinus fuscicollis to hours in which Callimico goeldii was alone. The proportion of feeding time devoted to insects did not differ significantly between hours associated vs. hours alone (F[1, 63] = 1.38, P < 0.25). Thus although Saguinus labiatus potentially flushes insects from the lower and middle canopies to Callimico goeldii in the understory, its proximity does not result in significant differences in the frequency of insectivory. In addition, there were no significant increases in the frequency of frugivory or mycophagy when Callimico goeldii was associated as compared to when it was alone. Thus, associations appear to result in a general increase in feeding behaviors, not an increase in consumption of any particular food type. There were seasonal differences, however, in Callimico goeldii feeding behaviors, with frugivory higher in the wet season (43%) than the dry season (19%) (F[1, 76] = 7.48, P < 0.01) and mycophagy higher in the dry season (25%) than the wet season (9%) (F[1, 76] = 4.73, P < 0.05). Although there were no significant changes in the frequency with which certain foods were consumed during polyspecific associations, feeding behaviors appear to be closely linked to polyspecific association rates. I therefore tested, through multiple regression analyses, whether changes in feeding behaviors during different months were good predictors of association rates. I used monthly dietary overlap (with Saguinus fuscicollis or S. labiatus), and the percentage of food types in feeding records (fruits, fungus, or insects), for the multiple regression analyses. Analyses showed that the percentage of feeding records that were on fruits (frugivory) was the best predictor of monthly association rates (F[5, 11] = 6.08, P < 0.05, P ≤ 0.05 to add, P ≥ 0.10 to remove (Fig. 4a)). It can also be seen that although frugivory was the best predictor, mycophagy was roughly inversely proportional to association rates (Fig. 4a), and monthly dietary overlap with both Saguinus species roughly parallel monthly association rates (Fig. 4b). DISCUSSION Groups of Saguinus fuscicollis and S. labiatus in this study, as in previous studies [Buchanan-Smith, 1990; Peres, 1992c; Buchanan-Smith, 1999], maintain associations exclusively with one group of the other species, sharing common home ranges, and defending common territories. In contrast, Callimico goeldii have a home range approximately six times larger than Saguinus, and move between different Saguinus groups throughout their range (Fig. 1). This is similar to association patterns described for Saimiri with Cebus apella [Terborgh, 1983; Podolosky, 1990], with one Saimiri group associating with multiple Cebus groups throughout its large home range. Associations were generally initiated by 154 / Porter Fig. 4. The mean percentage of time Callimico goeldii was in association with Saguinus troops, plotted with (a) the frequency of frugivory and mycophagy, and (b) dietary overlap values with Saguinus fuscicollis and S. labiatus. Callimico goeldii through contact calls that were responded to by both Saguinus species. Once together, however, Saguinus labiatus led group activities, such as travel and feeding, and initiated resting bouts. Associations between Callimico goeldii and Saguinus troops occurred much more frequently than expected by chance. On average, Callimico goeldii was found in association with Saguinus groups during 53% of observations. Monthly association rates varied, however, from 89% of observations in the wet-season month of February to 13% in the dry-season month of July. Multiple regression analyses showed that association frequency was best predicted by the frequency of fruit feeding: months of high frugivory were also the months of high association Callimico goeldii Polyspecific Associations / 155 rates (Fig. 4a). In addition, dietary overlap rates roughly parallel association rates. This suggests that polyspecific associations are linked closely with foraging compatibility and foraging benefits. Foraging compatibility has been proposed to constrain polyspecific associations among other primate species. Cords [1990a] found that Cercopithecus species associated more frequently at a site where they had higher dietary overlap than the site where dietary overlap was low. Furthermore, Chapman and Chapman  found that food availability and interspecific feeding competition limited associations between red colobus groups and groups of other primate species. The data from the present study suggest that differences in foraging and feeding strategies among Callimico goeldii, Saguinus fuscicollis, and S. labiatus when fruits are scarce reduces their compatibility with one another. During fruit scarcity, Saguinus species eat nectar and exudates while Callimico goeldii eats fungi [Porter, in press]. Fungi are eaten by Callimico goeldii throughout the year, particularly in the dry season, but are rarely eaten by Saguinus species [Porter, in press]. Fungi are patchy and ephemeral resources that are found more often in stream-edge and bamboo habitats [Hanson, 2000], habitats that appear to be less useful for Saguinus troops. Indeed, Callimico goeldii habitat use varied significantly with association status and season. Callimico goeldii used both stream-edge and bamboo habitats more while alone than while associated, and used stream-edge habitats more in the dry season, when mycophagy was highest. These patterns of habitat use suggest that foraging for fungi limits Callimico goeldii associations with Saguinus species that forage on other resources (exudates and nectar) found in other habitats [Porter, in press]. Callimico goeldii was found on average nearly 1 m higher during associations than alone. This increase in height use appears to allow Callimico goeldii to expand the area in which it forages and feeds, a benefit that is particularly important during the wet season, when feeding rates increased during polyspecific associations. During the wet season, fruits were the principle food resource for Callimico goeldii [Porter, 2000; in press], and 89% of these fruits were consumed above 5 m from the ground [Porter, 2000; in press)]. Thus it is during months when fruits are available that Callimico goeldii would benefit most from increasing height use. Although many studies have reported decreased scanning rates during associations, Callimico goeldii showed no reduction in scanning behavior if habitat type and height were controlled for in the analysis. Thus there is no support for the group effect: associations do not permit a decrease in scanning rates. However, overall predation risk is likely to be lower in an associated group due to decreased probability of capture [Roberts, 1996] and increased number and spread of vigilant individuals in a larger group [Pook & Pook, 1982; Peres, 1989; Peres, 1992b; Caine, 1993; Peres, 1993; Hardie & Buchanan-Smith, 1997]. Increased predator avoidance in the mixed group is likely to explain why Callimico goeldii leaves its preferred habitat (dense understory [Porter, 2000]) for more exposed habitats (primary forest and the lower and middle canopies) more frequently while in the presence of Saguinus groups. Indeed, this is supported by anecdotal observations that if Callimico goeldii arrived at a fruiting tree before nearby Saguinus groups, it would wait under the tree until Saguinus arrived before climbing into the canopy to feed (Porter and Hanson, personal observations). In addition, as Callimico goeldii follows Saguinus labiatus while in an associated group, it is possible that C. goeldii uses S. labiatus as a guide to fruit resources in the canopy, thereby increasing the height at which it forages and 156 / Porter feeds. Parasitic relationships, in which one species increases food consumption by following a species with greater fruit resource knowledge, have been demonstrated with captive tamarins [Prescott & Buchanan-Smith, 1999], with Saimiri sciureus and Cebus apella [Terborgh, 1983; Podolosky, 1990] and in associated Cercopithecus groups [Cords, 1990a]. In this study, although associations led to an increase in feeding during the wet season, they did not lead to an increase in frugivory (as percentage of diet). Thus, while Callimico goeldii may eat fruits regardless of association status, it can eat fruits more frequently when in an associated group. While associations improve fruit feeding and foraging, Callimico goeldii does not appear to gain from associations with Saguinus through improved insectivory. Studies show that orthopterans, the principal type of arthropods eaten by Callimico goeldii, Saguinus fuscicollis, and S. labiatus (41%, 58%, and 100% of insect records, respectively [Porter, in press]), are at very low abundance and sparsely distributed in tropical forests [Penny & Arias, 1982; Porter, in press]. Thus, Saguinus troops are not likely not to lead Callimico goeldii to orthopteran resources. Furthermore, although Saguinus labiatus could potentially flush insects to Callimico goeldii, given the differences in their foraging heights, there were no significant increases in insectivory for Callimico goeldii during polyspecific associations. Finally, traveling higher in the forest would not aid Callimico goeldii insect foraging, as insects were eaten almost exclusively in the understory. Improved fruit foraging, therefore, appears to be the principal benefit that Callimico goeldii gains from polyspecific associations. Given recent evidence suggesting that polyspecific associations can vary considerably over small spatial and temporal scales [Chapman & Chapman, 2000], it is important that further studies be conducted in order to compare association patterns from this study with those of Callimico goeldii in other areas. It is particularly important to more closely assess the role that Callimico goeldii, Saguinus labiatus, and S. fuscicollis have in predator detection in the understory and canopy. The high rates of scanning by Callimico goeldii, regardless of its association status, suggest that one reason Saguinus groups may actively maintain associations with Callimico goeldii is to increase their ability to avoid understory predators. ACKNOWLEDGMENTS Thanks to the Ministerio de Desarollo y Medio Ambiente of La Paz, the Colección Boliviana de Fauna, and the Herbario Nacional de Bolivia for assistance in obtaining my research permits. Special thanks to my guide, Edilio Nacimento B., and my research assistants, Kristin Donaldson, Laura Johnson, and Gonzalo Calderon V., for all their help in data collecting. Thanks to Drs. Patricia Wright, John Fleagle, Charles Janson, Diane Doran, and Anita Christen for their help in formulating this project and making comments on this manuscript. REFERENCES Boinski S. 1987. Birth synchrony in squirrel monkeys (Saimiri oerstedi). Behav Ecol Sociobiol 21:393–400. Buchanan-Smith H. 1990. Polyspecific association of two tamarin species, Saguinus labiatus and Saguinus fuscicollis, in Bolivia. Am J Primatol 22:205–214. Buchanan-Smith H. 1991. Field observations of Goeldi’s monkey, Callimico goeldii, in northern Bolivia. Folia Primatol 57:102–105. Buchanan-Smith H. 1999. Tamarin polyspecific associations: forest utilization and stability of mixed-species groups. Primates 40:233–247. Caine NG. 1984. Visual scanning by tamarins. Folia Primatol 43:59–67. Callimico goeldii Polyspecific Associations / 157 Caine N. 1993. Flexibility and co-operation as unifying themes in Saguinus social organization and behaviour: the role of predation pressure. In: Rylands A, editor. Marmosets and tamarins. New York: Oxford University Press. p 200–219. Chapman CA, Chapman CA, Chapman LJ. 1996. Mixed-species primate groups in the Kibale Forest: Ecological constraints on association. Int J Primatol 17:31–50. Chapman C, Chapman L. 2000. Interdemic variation in mixed-species association patterns: common diurnal primates of Kibale National Park, Uganda. Behav Ecol Sociobiol 47:129–139. Christen A, Geissmann T. 1994. A primate survey in northern Bolivia, with special reference to Goeldi’s monkey, Callimico goeldii. Int J Primatol 15:239–273. Cords M. 1990a. Mixed-species association of East African guenons: general patterns or specific examples? Am J Primatol 21: 101–114. Cords M. 1990b. Vigilance and mixed-species association of some East African forest monkeys. Behav Ecol Sociobiol 26:297–300. Ferrari S, Lopes Ferrari M. 1990. Predator avoidance behavior in the buffy-headed marmoset, Callithrix flaviceps. Primates 31:323–338. Fitzgibbon CD. 1990. Mixed-species grouping in Thomson’s and Grant’s gazelles: the antipredator benefits. Anim Behav 31: 1116–112. Garber PA. 1988. Diet, foraging patterns, and resource defense in a mixed species troop of Saguinus mystax and Saguinus fuscicollis in Amazonian Peru. Behaviour 105:18–34. Garber P. 1993. Feeding ecology and behavior of the genus Saguinus. In: Rylands AB, editor. Marmosets and tamarins. Oxford: Oxford University Press. p 273–295. Garber P. 2000. The behavioral ecology of mixed species troops of Callimico goeldii, Saguinus labiatus and S. fuscicollis in northwestern Brazil. Am J Phys Anthropol Suppl 30:155. Garber P, Bicca-Marques J. In preparation. Evidence of predator sensitive foraging in small and large-scale space in free-ranging tamarins (Saguinus fuscicollis, Saguinus imperator and Saguinus mystax). In: Miller L, editor. Eat or be eaten: predator sensitive foraging in primates. Cambridge: Cambridge University Press. Hanson A. 2000. Habitat use in relation to diet, with particular emphasis on mycophagy, by Callimico goeldii in Pando, Bolivia. Masters thesis, State University of New York at Stony Brook. Hardie S, Buchanan-Smith H. 1997. Vigilance in single- and mixed-species groups of tama- rins (Saguinus labiatus and Saguinus fuscicollis). Int J Primatol 18:217–234. Heymann EW. 1990. Interspecific relations in a mixed-species troop of moustached tamarins, Saguinus mystax, and saddleback tamarins, Saguinus fuscicollis (Platyrrhini: Callitrichidae), at the Rio Blanco, Peruvian Amazonia. Am J Primatol 21: 115–127. Hodge MA, Uets GW. 1996. Foraging advantages of mixed-species association between solitary and colonial orb-weaving spiders. Oecologia 107:578–587. Holenweg A-K, Nöe R, Schabel M. 1995. Waser’s gas model applied to associations between red colobus and diana monkeys in the Tai National Park, Ivory Coast. Folia Primatol 67:125–136. Isbell L. 1990. Sudden short-term increase in mortality of vervet monkeys (Cercopithecus aethiops) due to predation in Amboseli National Park, Kenya. Am J Primatol 21:41–52. Janson CH. 1988. Intra-specific food competition and primate social structure: a synthesis. Behaviour 105:1–17. Janson CH, Goldsmith ML. 1995. Predicting group size in primates: foraging costs and predation risks. Behav Ecol 6:326–336. Koenig A. 1998. Visual scanning by common marmosets (Callithrix jacchus): functional aspects and the special role of adult males. Primates 39:85–90. Landeau L, Terborgh J. 1986. Oddity and the “confusion effect” in predation. Anim Behavior 34:1372–1380. Lopes M, Ferrari S. 1994. Foraging behavior of a tamarin group (Saguinus fuscicollis weddeli) and interactions with marmosets (Callithrix emiliae). Int J Primatol 15:373–387. Martin P, Bateson P. 1993. Measuring behaviour: an introductory guide. Cambridge: Cambridge University Press. Milton K. 1980. The foraging strategy of howler monkeys: a study in primate economics. New York: Columbia University Press. Munn CA, Terborgh JW. 1979. Multi-species territoriality in neotropical foraging flocks. Condor 81:338–347. Nöe R, Bshary R. 1997. The formation of red colobus-diana monkey associations under predation from chimpanzees. Proc R Soc Lond B 264:253–259. Norconk M. 1990. Mechanisms promoting stability in mixed Saguinus mystax and S. fuscicollis troops. Am J Primatol 21: 159–170. Oates JF, Whitesides GH. 1990. Association between olive colobus (Procolobus verus), diana guenons (Cercopithecus diana), and 158 / Porter other forest monkeys in Sierra Leone. Am J Primatol 21:129–146. Penny N, Arias J. 1982. Insects of an Amazon forest. New York: Columbia University Press. Peres C. 1989. Costs and benefits of territorial defense in wild golden tamarins, Leontopithecus rosalia. Behav Ecol Sociobiol 25:227–233. Peres C. 1992a. Consequences of joint-territoriality in a mixed-species group of tamarin monkeys. Behavior 123:220–246. Peres CA. 1992b. Consequences of joint-territoriality in a mixed-species group of tamarin monkeys. Behavior 123:220–246. Peres CA. 1992c. Prey-capture benefits in a mixed-species group of Amazonian tamarins, Saguinus fuscicollis and S. mystax. Behav Ecol Sociobiol 31:339–347. Peres CA. 1993. Anti-predation benefits in a mixed-species group of Amazonian tamarins. Folia Primatol 61:61–76. Peres C. 2000. Territorial defense and the ecology of group movements in small-bodied neotropical primates. In: Boinski S, Garber P, editors. On the move: how and why animals travel in groups. Chicago: Chicago University Press. p 100–124. Podolosky RD. 1990. Effects of mixed-species association on resource use by Saimiri sciureus and Cebus apella. Am J Primatol 21:147–158. Pook A, Pook G. 1982. Polyspecific association between Saguinus fuscicollis, Saguinus labiatus, Callimico goeldii and other primates in north-western Bolivia. Folia Primatol 38:196–216. Popp JW. 1988. Scanning behavior of finches in mixed-species groups. Condor 90:510–511. Porter L. 2000. Behavior and ecology of the Goeldi’s monkey (Callimico goeldii) in northern Bolivia. PhD thesis, State University of New York at Stony Brook. Porter L. In press. Dietary differences among sympatric callitrichines in northern Bolivia: Callimico goeldii, Saguinus fuscicollis and Saguinus labiatus. Int J Primatol. Prescott MJ, Buchanan-Smith HM. 1999. Foraging efficiency in single- and mixedspecies troops of tamarins. Folia Primatol 70:221–231. Pulliam H. 1973. On the advantages of flocking. J Theor Biol 38:419–422. Roberts G. 1996. Why individual vigilance declines as group size increases. Anim Behav 51:1077–1086. Savage A, Snowdon C, Giraldo L, Soto L. 1996. Parental care patterns and vigilance in wild cotton-top tamarins (Saguinus oedipus). In: Norconk M, Rosenberger A, Garber P, editors. Adaptive radiations of neotropical primates. New York: Plenum Press. Sokal RR, Rohlf FJ. 1995. Biometry, 3rd edition. New York: WH Freeman. 887 p. Terborgh J. 1983. Five New World primates. Princeton: Princeton University Press. Terborgh J. 1990. Mixed flocks and polyspecific associations: costs and benefits of mixed groups to birds and monkeys. Am J Primatol 21:87–100. Treves A. 1998. The influence of group size and neighbors on vigilance in two species of arboreal monkeys. Behaviour 135: 453–481. van Schaik CP, van Noordwijk M. 1986. The hidden costs of sociality: intra-group variation in feeding strategies in Sumatran long-tailed macaques (Macaca fascicularis). Behaviour 99:296–315. Wachter B, Schabel M, Nöe R. 1997. Diet overlap and poly-specific associations of red colobus and diana monkeys in the Tai National Park, Ivory Coast. Ethology 103:514–526. Waser P. 1982. Primate polyspecific associations: do they occur by chance? Anim Behav 30:1–8. Waser PM. 1984. “Chance” and mixed-species associations. Behav Ecol Sociobiol 15:197–202. Whitesides GH. 1989. Interspecific associations of diana monkeys, Cercopithecus diana, in Sierra Leone, West Africa: biological significance or chance? Animal Behav 37:760–776. Wolda H. 1981. Similarity indices, sample sizes and diversity. Oecologia 50:296–302. Wright P. 1998. Impact of predation risk on the behavior of Propithecus diadema edwardsi in the rain forest of Madagascar. Behaviour 135:483–512. Yoneda M. 1981. Ecological studies of Saguinus fuscicollis and Saguinus labiatus with reference to habitat segregation and height preference: Overseas Research Reports of New World monkeys. Primate Research Institute, Kyoto University, Kyoto, Japan. p 43–50. Yoneda M. 1984. Comparative studies on vertical separation, foraging behavior and traveling mode of saddle-backed tamarins (Saguinus fuscicollis) and red-chested moustached tamarins (Saguinus labiatus) in northern Bolivia. Primates 25:414–422.